Building Your AI Agent

Kazi's AI Agents let you create specialized assistants that understand your business context. Whether you need a proposal writer, data analyst, or client communication assistant, you can build one in minutes.

Creating a New Agent

Navigate to AI Agents in the sidebar and click Create Agent. Give your agent a name, description, and choose an avatar. The description helps Kazi understand the agent's purpose and provide better suggestions.

Choosing a Model

Select from 12 AI models including Claude Opus, Claude Sonnet, GPT-5, Gemini 3 Flash, and more. Each model has different strengths — Claude excels at nuanced writing, GPT at code generation, and Gemini at fast responses. Start with the default (Gemini 3 Flash) for cost efficiency, then upgrade if you need more capability.

Configuring the System Prompt

The system prompt defines your agent's personality, expertise, and constraints. Be specific: instead of "You are a helpful assistant," try "You are an expert freelance proposal writer who creates compelling, conversion-focused proposals for design agencies. Always include a timeline, deliverables section, and three pricing tiers."

Using Agent Tools

Agents can access your Kazi data through built-in tools — looking up invoices, searching clients, checking project status, and analyzing financials. Enable the tools relevant to your agent's purpose. For example, a financial analyst agent should have access to invoice lookup and financial summary tools.

Memory and Context

Each agent maintains conversation memory, so it remembers past interactions and learns your preferences over time. You can also feed it context documents like brand guidelines, rate cards, or project briefs to make its responses more relevant.